A critical challenge for psychiatry is determining the best way to characterize psychopathology given the continuously distributed and highly correlated nature of most psychological symptoms. An increasing body of evidence on the structure of mental illness indicates the presence of a latent general factor of psychopathology (GF), also referred to as a P-factor, that explains a significant amount of the common variance in the expression of multiple psychiatric symptoms and disorders. The presence of a GF has substantial implications for psychopathology research and treatment because it suggests that a comprehensive understanding of mental illness requires an untangling of nonspecific risk factors from narrower, dimension-specific, risk factors or features. Caspi et al., (2014) suggest that schizophrenia symptoms are so highly correlated with the GF that they are mostly an expression of the GF. However, methodological limitations, such as the use of community and youth samples with few psychotic patients, restrict conclusions from the existing literature. We propose to test the hypothesis that schizophrenia spectrum and psychotic symptoms load heavily on both a GF AND a separate higher-order psychosis factor in a sample of 800 adult psychiatric and medical treatment seeking subjects that includes a substantial group of patients with psychotic disorders as well as patients with significant externalizing and internalizing symptoms. The potential utility of defining an accurate structural model of severe psychopathology lies in its ability to provide more accurate dimensional phenotyping than can be achieved by categorical diagnoses of primary and comorbid conditions. This may in turn aid in understanding mechanisms involved in the etiology and expression of illness as well as predicting its course and optimal treatment. Toward this end, we aim to determine the extent to which patient factor scores at different levels of the structure of psychopathology (GF, 2nd-order psychosis and narrower 1st order symptom dimensions) predict neural and neuropsychological features that have previously been associated with schizophrenia but have never been tested while statistically controlling for dimensional features of psychopathology outside of the schizophrenia spectrum. Using this quantitative approach in a subsample of 300 participants, we will explicitly test the hypothesis that some neural and cognitive correlates of schizophrenia, such as the volume of the anterior cingulate area, are nonspecific correlates of the GF, while others, such as temporal cortical sensory processing abnormalities, are specific to a 2nd-order psychosis factor, and will show associations that remain significant even after controlling for the GF. In order to assess the prognostic significance of the GF, we will test whether scores on the GF are predictive of the course of illness in 150 patients experiencing first-episode psychosis (even after controlling for previously defined prognostic indicators). Taken together, the study will provide the most comprehensive test to date of the relevance of GF model to understanding the expression and neural correlates of severe psychopathology.